1. Field of the Invention
This invention relates to adaptive optics systems used to compensate for rapidly changing air turbulence, and to spatial light modulator designs and methods used therein.
2. Description of the Related Art
Innovative adaptive optics systems are a class of adaptive optics systems used to compensate for atmospherically-induced aberrations that are imposed upon a beam traveling through the atmosphere to a detector. Such systems generally transmit a reference laser beam to a target which is being observed. The laser beam is reflected from the target and accompanies the target image beam back to the detector, picking up the same aberrations along the way. At the detector the reflector laser beam is beat against an unaberrated reference laser beam of the same frequency to produce a spatial interference pattern that corresponds to the spatial phase distortions imposed upon the input target beam. A control beam that bears the interference pattern as a spatial intensity distribution is applied to a spatial light monitor (SLM), which translates the pixelized intensity pattern into a pixelized phase adjustment for the aberrated input beam. The phase adjustment provided by the control beam compensates for the atmospheric distortions, and allows an essentially undistorted output beam to be obtained.
Both innovative and conventional adaptive optics systems have not been very successful for applications involving rapidly changing atmospheric conditions, such as imaging through boundary layer turbulence from an aircraft platform, atmospheric turbulence, and astronomical studies under high velocity wind conditions. In one approach a liquid crystal light valve (LCLV) has been used for the SLM. This approach is described in U.S. Pat. No. 5,090,795 to T. R. O'Meara and C. C. Valley, assigned to Hughes Aircraft Company, the assignee of the present invention. The single layer LCLV used as a spatial light modulator in this system had a maximum closed-loop response speed of 1 millisecond or less, whereas much faster response times are required for rapidly changing atmospheric conditions. Attempts to speed up the LCLV response time by multi-pass designs were found to sacrifice optical efficiency and increase the difficulty of alignment.
Membrane light modulators have also been developed, in which a spatial optical intensity pattern is converted to a pixelized electron flow that accumulates charges on the pins of a charge transfer plate. Such light modulators and their component are described in U.S. Pat. Nos. 4,800,263 and 4,822,993 to Dillon et al., and U.S. Pat. Nos. 4,794,296 and 4,863,759 to Warde et al. The accumulated charges produce pixelized deformations in a deformable membrane mirror from which the aberrated input beam is reflected. Since each mirror deformation imparts a phase shift to its corresponding input beam pixel in proportion to the amount of deformation, which in turn is a function of the accumulated charge on the corresponding charge transfer pin, the interference intensity pattern produced by beating the aberrated and unaberrated reference laser beams against each other is translated to a pixelized phase correction for the input beam. Since charge continuously accumulates on the plate, the prior systems are operated in a frame mode, in which input frames are separated by erase pauses during which the accumulated charge is removed from the charge transfer plate. The result is a frame rate of typically 10 milliseconds or slower. Furthermore, the charge removal process causes the mirror state to be unsuitable for compensation purposes for appreciable intervals, such that the beam to be compensated would typically be required to be shut off during erase and during a portion of the subsequent reset period.
In FIG. 2 of U.S. Pat. No. 4,794,296 a low resistance enhancer coating is formed over the back surface of the charge transfer device so that ". . . electronic charges bleed off the enhancer coating at a specific rate"; this is said to allow the subject charge transfer signal processor to be operated in a continuous mode. However, the continuous low resistance coating would establish different charge drain rates for the device's different charge transfer pins, depending upon each pin's particular location. Some pins would see a longer and thus higher resistance path, while others would see a shorter and thus lower resistance path. The varying charge drainage rates result in different response times for the different pin locations, making the device unusable for applications such as adaptive optics that require rapid response rates.
The prior charge transfer plates were fabricated by forming a matrix of holes through an insulative substrate, and then filling the holes with a conductive material. This fabrication process often leaves pathways through the charge transfer plate, through which outgassed organic materials from the membrane mirror can seep to contact the photocathode used to generate the pixelized electron flow onto the charge transfer plate. The photocathode is highly sensitive to such materials, and rapidly deteriorates in their presence. Also, the actual process of forming the conducting pin matrix can trap gases in the holes, which later outgas and again reduce the lifetime of the photocathode. The difference of expansion coefficients between the pins and glass also makes a thermally unstable system which limits the types of microchannel plate amplifiers that can be used in the device.
A further problem with prior membrane light modulators relates to the manner in which the membrane mirrors are deformed by the accumulated charges on the charge transfer plate. Each pixel of the membrane mirror deforms along a curved surface, which imparts a similar curvature to the wavefront of the input beam pixel which it reflects. Since different mirror pixels will be deformed by differing amounts, and will therefore have different curvatures, the reflected beam pixels will similarly have differing wavefront curvatures and focal lengths. However, it is highly desirable that the reflected wavefronts be flat. Since only the centers of each mirror pixel are relatively flat, only the light reflected from these small center areas is used. In prior applications the rest of the light, typically representing in excess of 99% of the total input beam, is simply discarded.